AI generated art

[Paul Meccano]

Your description still strikes me that the AI has a catalog of "John Berkley spaceships" to work from, otherwise it wouldn't be able to produce spaceships.

Exactly that.
In the last few days Meta had released an AI addition to messenger. Click and agree at your peril.
It allows them to use anything you post over all Meta platforms. Meta, like all of the most aggressive fishing companies do, have you down as agreeing until you write to tell them you don’t.

I love extreme technology, just as much I hate it. And for that reason I don’t get involved in AI. Not while I don’t get a choice.

 

[Mon0Zer0]

It interests me for a time and then gets kind of boring. But then I see one that surprises me.

Those fake trailers--I saw this one for a Snow White movie.
The animation looks great but the composition and content action is just so repetitive and sterile.
Painful painful narration:

On the other hand
this one is pretty darn good in terms of feeling authentically retro:

I'm skeptical whether AI video will ever get to a stage where whole performances can be created believably and have as much emotional weight as real people, but I think using these as filters for CG or live action could be a great use. Assets built, but not textured, in CG and then "skinned" using AI. So you have a conventionally shot movie which is augmented to look like the 50s / 60s snow white with the same kind of lenses, film stock and colour grading, skin texture, makeup but with the kind of FX that couldn't be dreamed of back then.

This is using Flux - a locally installed AI which is a tonne better than Stable Diffusion and doesn't look so AI-ish.

 

[Cthulhu.Science]

I'm skeptical whether AI video will ever get to a stage where whole performances can be created believably and have as much emotional weight as real people, but I think using these as filters for CG or live action could be a great use. Assets built, but not textured, in CG and then "skinned" using AI. So you have a conventionally shot movie which is augmented to look like the 50s / 60s snow white with the same kind of lenses, film stock and colour grading, skin texture, makeup but with the kind of FX that couldn't be dreamed of back then.

This is using Flux - a locally installed AI which is a tonne better than Stable Diffusion and doesn't look so AI-ish.


View attachment 128678

View attachment 128679

View attachment 128680

Where does one download Flux???

Does everything look digital art-ish?

 

[Mon0Zer0]

Where does one download Flux???

Does everything look digital art-ish?

Do you have an Nvidia RTX series graphic card with at least 12GB of ram? Unless you do you won't be able to get it to run. I have an RTX3060 12GB and it just about runs (approx. 2-3 mins per image compared to Stable Diffusion!) You need to run it using a front end like Automatic 1111 or Comfyui.

Flux mainly excels at photorealistic imagery. Digital Art is what it's worst at!

 

[Cthulhu.Science]

Do you have an Nvidia RTX series graphic card with at least 12GB of ram? Unless you do you won't be able to get it to run. I have an RTX3060 12GB and it just about runs (approx. 2-3 mins per image compared to Stable Diffusion!) You need to run it using a front end like Automatic 1111 or Comfyui.

Flux mainly excels at photorealistic imagery. Digital Art is what it's worst at!

As it turns out. I have the same card that you do.
I just double checked the RAM on mine in case there were variants. There may be variants but mine is 12 GB also.

I just searched Flux and google offered me about 20 pages of online image generators using Flux, but I still haven't found the site to download it from.
Do I just try to search the Microsoft store?? I did that and got 150,000000000 variants of apps using Flux but I'm not sure how to get to the basic item.

Any guidance?

 

[Mon0Zer0]

Try this:

 

[Cthulhu.Science]

I did get a pretty good girl with a pearl earring with Stable Diffusion

 

[Cthulhu.Science]

Try this:

Thanks!

 

[PETERTHEKNOTTER]

I don't think so. People don't generally read a lot of Balzac and then write a series of Balzac-like things.


The AI isn't trained to think like Berkley or to have a rationalization of technology like Berkley. It is being trained to reassemble visual cues around certain "spaceship-like rules", which have more to do with some rules about up and down, symmetricality and perspective combined with "pointy to the left, holes to the right" guidance. It just doesn't seem much different than a text generator.

histoires droles... has given me much inspiration.. and, a bit like the glass bead game by hesse, creates a synthesis of much diversity in an organic way which the viewer may often, given a certain level or capacity to observe, may taste a hint or flavour of a balzac or asimov, baudelaire or heinlein, pirsig or poe?